Comparison of [18F]DCFPyL and [68Ga]Ga-PSMA-HBED-CC for PSMA-PET Imaging in Patients with Relapsed Prostate Cancer

Purpose Gallium-68 (Ga-68)-labeled tracers for imaging expression of the prostate-specific membrane antigen (PSMA) such as the [68Ga]Ga-PSMA-HBED-CC have already demonstrated high potential for the detection of recurrent prostate cancer. However, compared to Ga-68, a labeling with fluorine-18 (F-18) would offer advantages with respect to availability, production amount, and image resolution. [18F]DCFPyL is a promising F-18-labeled candidate for PSMA-positron emission tomography (PET) imaging that has been recently introduced. In the current study, we aimed to compare [68Ga]Ga-PSMA-HBED-CC and [18F]DCFPyL for clinical use in biochemically relapsed prostate cancer. Procedures In 14 selected patients with PSA relapse of prostate cancer, [18F]DCFPyL PET/X-ray computed tomography (CT) was performed in addition to [68Ga]Ga-PSMA-HBED-CC PET/CT. A systematic comparison was carried out between results obtained with both tracers with regard to the number of detected PSMA-positive lesions, the standardized uptake value (SUV)max and the lesion to background ratios. Results All suspicious lesions identified by [68Ga]Ga-PSMA-HBED-CC were also detected with [18F]DCFPyL. In three patients, additional lesions were observed using [18F]DCFPyL PET/CT. The mean SUVmax in the concordant [18F]DCFPyL PSMA-positive lesions was significantly higher as compared to [68Ga]Ga-PSMA-HBED-CC (14.5 vs. 12.2, p = 0.028, n = 15). The mean tumor to background ratios (n = 15) were significantly higher for [18F]DCFPyL compared to [68Ga]Ga-PSMA-HBED-CC using kidney, spleen, or parotid as reference organs (p = 0.006, p = 0.002, p = 0.008), but no significant differences were found using the liver (p = 0.167) or the mediastinum (p = 0.363) as reference organs. Conclusion [18F]DCFPyL PET/CT provided a high image quality and visualized small prostate lesions with excellent sensitivity. [18F]DCFPyL represents a highly promising alternative to [68Ga]Ga-PSMA-HBED-CC for PSMA-PET/CT imaging in relapsed prostate cancer.


Introduction
T he prostate-specific membrane antigen (PSMA) is overexpressed on the cell surface of prostate cancer (PC) cells [1]. Recent studies with gallium-68 (Ga-68)labeled Glu-NH-CO-NH-Lys-(Ahx) ([ 68 Ga]Ga-PSMA-HBED-CC) have shown the potential of this radioligand to detect relapses and metastases of PC with improved contrast when compared to [ 18 F]fluoromethylcholine positron emission tomography (PET)/X-ray computed tomography (CT) [2][3][4][5]. Additionally, iodine-124, iodine-131, and lutetium-177-labeled PSMA ligands have been reported for dosimetric and therapeutic use [6,7]. Several PSMA inhibitors were developed which compromise different pharmacophoric structures to interact with the binding pocket for N-acetyl-L-aspartyl-L-glutamate. PSMA-binding ligands are bound to the extracellular domain of PSMA. However, the transmembranous location of the binding domain and its enzyme activity enable the subsequent internalization of these ligands. The prostate-specific membrane antigen is also expressed in the tumor-associated neovasculature of gastric and colorectal cancer [8,9]. Thus, these PSMAselective ligands may be of interest for imaging other tumor types in the future.
In comparison to currently broadly applied Ga-68labeled PSMA ligands, fluorine-18 (F-18)-labeled compounds would offer some important advantages. This includes not only an increase of the number of examinations owing to the higher production capacity but also an excellent image quality. The latter will be a result of optimized tracer doses leading to high imaging statistics and the decay properties of F-18 itself. F-18 exhibits a low positron emission energy of 0.6 MeV. Therefore, the distance to decelerate the positron in human tissue is much shorter in comparison to Ga-68 (β + -energy=2.3 MeV) resulting in a much higher image resolution. Recently, Chen and colleagues [10] have published first data on [ 18 F]DCFPyL (2-(3-{1-carboxy-5-[(6-[ 18 F]fluoro-pyridine-3-carbonyl)-amino]-pentyl}-ureido)-pentanedioic acid), a new PSMA-selective ligand with a high binding affinity for PSMA. It was suggested that this compound may represent a highly promising candidate for PET imaging of PSMA-overexpressing tissues [10].
In this work, a comparison between [ 68 Ga]Ga-PSMA-HBED-CC and [ 18 F]DCFPyL PET/CT in patients with a PSA-relapse of prostate cancer was carried out.
We assumed that [ 18 F]DCFPyL would bear the potential of high diagnostic accuracy due to the following: Availability of higher tracer activity amounts in combination with the longer half-life of F-18, enabling imaging at later time points resulting in higher clearance and lower nonspecific binding Higher image quality due to lower positron emission energy of F-18 We aimed to confirm and potentially extend the findings obtained with our diagnostic standard, [ 68 Ga]Ga-PSMA-HBED-CC PET/CT, by adding a second scan with [ 18 F]DCFPyL, representing a different PSMA tracer. By choosing this approach, we tried to collect accumulating evidence in the diagnostic assessment of selected patients and to compare the properties of [ 18 F]DCFPyL with [ 68 Ga]Ga-PSMA-HBED-CC at the same time.

Patient Characteristics
In this study, 14 selected patients were included who underwent both [ 68 Ga]Ga-PSMA-HBED-CC PET/CT and [ 18 F]DCFPyL PET/ CT with the aim to determine if one of the two tracers exhibits better detection of recurrent prostate cancer and/or metastases. The patients had biochemical relapse of prostate cancer after initial curative treatment, either radical prostatectomy or radiation-based therapy. All patients showed rising PSA level and suspected progressive disease following prior treatment of prostate cancer. In 11 patients, the PSA level had increased to more than 1 ng/ml. All patients underwent the examination as part of their clinical workup. Patients were selected for the dual-scan procedure assuming that a thorough diagnostic assessment would have a significant influence on their individual subsequent therapeutic measures, i.e., local versus systemic therapy, indication for or exclusion from surgical treatment or external beam radiation therapy. This included, e.g., suspected false-negative or equivocal PET/CT results, detection of a solitary metastasis, or oligometastatic status potentially accessible for local therapy.
First, [ 68 Ga]Ga-PSMA-HBED-CC PET/CT was performed, representing the standard procedure at our center. Subsequently, in 14 selected patients, additional [ 18 F]DCFPyL PET/CT was carried out within a period of 3 weeks following the first scan. All patients underwent the examinations as part of the clinical workup in order to accumulate diagnostic evidence and potentially optimize their individual treatment. This study does not represent a systematic clinical trial, comparing the clinical value of two diagnostic instruments in a blinded fashion. All patients signed informed consent regarding the scientific evaluation of their data.
The PSMA enzyme inhibition potency of compound 3 was determined with a modified Amplex Red glutamic acid assay after incubation with the cell lysates of LNCaP cell extracts in the presence of NAAG for 2 h at 37°C. The enzyme inhibitory constant (Ki) for compound 3 was 1.1±0.1 nmol/l, comparable with that of ZJ-43, which was 1.4±0.2 nmol/l under same measurement conditions. ZJ-43 is a urea-based potent inhibitor of NAAG and is used as an internal reference in the assay.

Imaging
The mean dosage of 128.  [14]. Attenuation correction was performed using the low-dose non-enhanced CT data.
To ensure comparability between different PET/CT scanners, reconstruction was performed via an OSEM algorithm (4 iterations and 14 subsets), followed by an intrinsic 5-mm Gaussian filter in all directions for the Biograph 16 True Point (Siemens Medical Solutions), containing full-ring dedicated PET and 16-slice CT instrumentation. When using the Biograph mCT Flow-Edge 128 PET/CT system (Siemens Medical Solutions) with a 128-slice spiral CT, iterative reconstruction was performed using 4 iterations and 12 subsets by an intrinsic 5-mm Gaussian filter.

Image Analysis
Image analysis was performed using an appropriate workstation and software (Syngo TrueD, Siemens, Erlangen, Germany). Clinical PET/CT reading was performed by three experienced specialists from the department of nuclear medicine and from the department of radiology in consensus in a side-by-side analysis of the results obtained with both tracers. Lesions were visually interpreted as suspicious for local relapse, lymph node metastasis, bone metastasis, or visceral metastasis.
The standardized uptake value (SUV) max was measured in up to three hottest lesions (as identified in the [ 68 Ga]Ga-PSMA-HBED-CC scan) and their counterpart in the [ 18 F]DCFPyL scan. Background SUV mean values were measured in a volume of interest (VOI) with 2 cm diameter in the liver, spleen, kidney, mediastinum, and parotid in all patients. For the calculation of mean values and to compare the SUV max values of the lesions and their ratios, SPSS 22 was used.

Results
Patient characteristics are shown in Table 1.

Verification of Lesions
The bone metastases in the lumbar vertebra L2 (patient 2) and in the thoracal vertebral Th3 (patient 12) were confirmed by CT retrospectively. In patient 7, a concordant left iliac PSMA-positive lymph node was detected with both tracers and an additional F-18 PSMApositive pelvic lymph was reported. This patient underwent surgical lymph node resection, and the left iliac metastasis was confirmed histologically. The additional pelvic metastasis was not detected by salvage lymph node dissection, but PSA remained increased postoperatively, indicating that this lesion may have represented the remaining PSA-producing tissue. In 1 patient, the solitary PSMA spot in the irradiated prostate, seen in the [ 6 8 Ga]Ga-PSMA-HBED-CC PET/CT and in the [ 18 F]DCFPyL PET/CT, was not confirmed by core biopsy (patient 4). In 1 patient, the concordant solitary PSMA-avid lymph node metastasis was confirmed histologically by systematic lymphadenectomy, while core biopsy was negative for local relapse despite a focal PSMA concentration within the irradiated prostate (patient 1). Examples of the PSMA-positive lesions are shown in Figs. 1, 2, 3, and 4.

Quantitative Comparison
The To compare the tumor/background contrast between the two tracers, we calculated SUV values in suspicious lesions/ background in seven patients with maximum three lesions, detectable with both PET procedures. This resulted in a comparison of altogether 15 tumor/background ratios. Most of the suspected lymph node/bone metastases showed a higher SUV ratio for [ 18 F]DCFPyL using the kidney (Fig. 5a) or the spleen as a reference region, but no significant difference was found when using for the liver (Fig. 5b) or the mediastinum as reference organ. In detail, the mean tumor to background ratios between the SUV max in the PSMA-avid lesions with [ 18

Clinical Consequences
In the majority of cases, [ 68 Ga]Ga-PSMA-HBED-CC and [ 18 F]DCFPyL PET/CT resulted in accumulating evidence, supporting further treatment decisions consistently. Both imaging series accordingly supported decisions towards core biopsy of the irradiated prostate in three patients, to local therapy options (lymph node dissection or radiotherapy) in four patients with one to two PSMA-positive lymph nodes and to systematic therapy in three patients. In four patients, neither imaging test detected a lesion explaining the rise in PSA levels. [ 18 F]DCFPyL PET/CT resulted in the detection of additional vertebral bone metastases in two patients (patient 2 and 12) in an oligometastatic state (patient 2 showed bone metastases also in [ 68 Ga]Ga-PSMA-HBED-CC PET/CT but to a lesser extent). This finding did not have immediate therapeutical consequences because both patients were asymptomatic at the time of their PET/CT examination. However, more accurate knowledge on the extent of bone metastases bears the potential to perform better targeted radiotherapy in the future (i.e., if the patients become symptomatic from their bone metastases). The detection of many PSMA-positive supradiaphragmatic lymph nodes with [ 18 F]DCFPyL in addition to retroperitoneal lymph node metastases also detected with [ 68 Ga]Ga-PSMA-HBED-CC supported the decision to omit local radiotherapy of the retroperitoneal lymphatic pathways (patient 12). In one patient, a second PSMA-avid lymph node in the upper pelvis was detected with [ 18 F]DCFPyL in addition to another pelvic lymph node (detected with both tracers). In this patient, systematic bilateral pelvic lymph node dissection was performed and not influenced by the [ 18 F]DCFPyL findings. Unfortunately, only the larger metastasis (positive with both tracers) was detected intraoperatively and the patient remained PSA positive after surgery (patient 7).

Discussion
We analyzed 14 patients who underwent both [ 68 Ga]Ga-PSMA-HBED-CC PET/CT and [ 18 F]DCFPyL PET/CT. In all cases, a rising PSA level indicated disease recurrence following prior treatment of prostate cancer.
The following findings emerge from this analysis:  One explanation for the higher detection rate using [ 18 F]DCFPyL can be found in the higher injected dose, which allowed later acquisition times, potentially leading to better signal to noise ratios due to reduction of nonspecific signal. Lower background activity has been observed for the [ 18 F]DCFPyL in some organs such as the kidney. Independently from injected dose, faster clearance of the tracer from non-target tissue and higher affinity, may potentially have contributed to the detection of additional skeletal metastases observed for [ 18 F]DCFPyL as compared to [ 68 Ga]Ga-PSMA-HBED-CC.
It is important to notice that the higher detection rate for PSMA-avid lymph nodes by [ 18 F]DCFPyL was not attributed to retroperitoneal PSMA foci, where the differentiation against celiac ganglia (i.e., false-positive findings) might be a question [15].
Radiolabeling with Ga-68 is an excellent alternative for imaging centers with expertise in handling the commercially available 68 Ge/ 68 Ga radionuclide generator but without own (cost-intense) cyclotron. However, a single preparation (0.3-0.6 GBq Ga-68) may only allow scanning up to two to four patients and the output depends on the half-life of the generator. As now established at our center, a single radiolabeling procedure of [ 18 F]DCFPyL resulted in a batch containing 6 to 7 GBq of [ 18 F]DCFPyL. This rendered it possible to examine six patients after a single preparation with an optimal dose, which represents an advantage for centers with an own cyclotron and a production of F-18 in the daily management. Furthermore, due to the longer halflife of F-18 (110 vs. 68 min for Ga-68), it may also allow transportation to remote sites from commercial vendors.
A limitation of the current study is the lack of a consistent histological validation of [ 18  Another limitation can be found in the selection of patients. We did not recruit patients for both imaging series prospectively, and the selected population cannot serve as a representative sample. Due to the selection of patients with complex clinical questions (several of them being negative in [ 68 Ga]Ga-PSMA-HBED-CC PET/CT), it might be possible that the diagnostic accuracy of [ 68 Ga]Ga-PSMA-HBED-CC PET/CT is underestimated.
Comparison of the two tracers has not been carried out under identical conditions, i.e., mean injected dose and start of acquisition have been higher respectively later for [ 18 F]DCFPyL, as compared to [ 68 Ga]Ga-PSMA-HBED-CC PET/CT. It cannot be excluded that the performance of [ 68 Ga]Ga-PSMA-HBED-CC PET/CT would have been more similar to [ 18 F]DCFPyL under more identical examination conditions. However, it was the aim of this study to compare the two tracers under the circumstances usually available in clinical application. Generally, lower doses and earlier acquisition periods p.i. are selected for 68 Ga-labeled tracers. The time period of 1 h between injection of [ 68 Ga]Ga-PSMA-HBED-CC and the start of PET/CT was reported by many publications [2][3][4][5].
It was beyond the scope of an observational study to show time-activity curves. The time period of 2 h between the injection of [ 18 F]DCFPyL and the start of data acquisition was chosen in the light of conventional receptor imaging in nuclear medicine using other tracers with contrast enhancement on delayed images and is in accord with the data including timeactivity curves published by Szabo and colleagues [14]. Szabo [16,17] might be responsible for false-negative PET/CT results in some patients. We have learned from other studies that imaging of the choline transport and phosphorylation may detect PSMA-negative metastases in some cases [2,3]. Future studies should investigate the diagnostic accuracy of F-18-or Ga-68labeled PSMA in the early PSA-relapse G1 ng/ml, when [ 11 C] or [ 18 F] choline have low detection rates [18,19] and should redefine the therapeutic impact of PET/CT [20][21][22]. Improved diagnostic accuracy should be confirmed by systematic comparison of sensitivity and specificity of both agents.

Conclusion
Our results indicate that the [ 18 F]-labeled compound [ 18 F]DCFPyL is a highly promising alternative to [ 68 Ga]-Ga-PSMA-HBED-CC for PSMA-PET/CT imaging in relapsed prostate cancer. Based on significantly higher SUV values in the PSMA-avid lesions, [ 18 F]DCFPyL PET/CT may represent a valuable tool to detect small prostate cancer lesions with high sensitivity.
Conflict of Interest. The authors declare that they have no competing interests.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.